Empirical Economics II (Empirische Wirtschaftsforschung II)

Welcome to the OPAL Learning Environment for the course "Empirical Economics II" in the summer term of 2024!

Empirical economics involves the application and development of methods for the empirical verification of economic theories or hypotheses.

I have compiled a summary of the most important R codes, which you can find in the "Exam Preparation" folder  (filename: "Important R Codes"). I hope you find it helpful for your exam preparation. Please do not hesitate to contact me if you have any questions. In case I update the file, I will post it here.

There is a new post regarding the practice exam, under "Posts and Updates" (09.07.2024, Updated 10.07.2024).

Dear students,

I have received feedback that we have reached the critical threshold of five evaluations. Thank you very much for your support!

I would be very grateful if you could participate in the "Lehrevaluation" of this course.

Here is the link: https://evaluation.tu-chemnitz.de/evasys/online.php?p=S91HQ

And here is the corresponding QR Code:

Thank you very much!

IMPORTANT INFORMATION

Time: Monday (weekly) 5:15-6:45 PM

Room: C25.031 (prev.: 2/W031)

Kick-off event: 08.04.2024, 5:15 PM

The next tutorial will take place on the 18th of June (7:30 AM - 9:00 AM).

The lecture on the 24th of June will be given online.

See the post from 28.05.2024 for more details.

LINKS ETC.

Link to shared Google Docs document (see post of 12.03.):

https://docs.google.com/document/d/1UnJBVpabqzfae-hHe0yEw4mKtFtuYoABq4nOBzxHxQA/edit?usp=sharing

POSTS AND UPDATES

17.07.2024 

Dear students,

I have compiled a summary of the most important R codes, which you can find in the "Exam Preparation" folder  (filename: "Important R Codes"). I hope you find it helpful for your exam preparation. Please do not hesitate to contact me if you have any questions. In case I update the file, I will post it here.

09.07.2024 (Updated on 10.07.2024):

Dear students,

Yesterday, we worked on the practice exam in class. You can find the PDF file with the exercises in the "Exam Preparation" folder. Please do not hesitate to contact me if you have any questions regarding the practice exam, especially those who were unable to attend yesterday.

In the context of the R-Exercise, we reviewed (among others) the interpretation of the coefficients in the linear probability model and the interpretation of marginal effects of probit and logit models. To review the interpretations, refer to the examples we discussed in class.

You can find the R-Code in the shared Google Docs document.

We also reviewed some of the examples discussed in class in the context of IV estimation.

Please check the Opal learning environment from time to time in case I post any updates.

I wish you good luck on the exam and a nice non-lecture period!

Best regards,

Linda Glawe

05.07.2024:

Dear students,

In the last lecture, someone asked whether the order of the arguments in the phtest() function matters. I checked this in R, and found that phtest(plm.fe, plm.re) and phtest(plm.re, plm.fe) produce the same output.

06.06.2024:

Dear students,

A quick note on the last question of the IV/TSLS quiz that we discussed on Monday.

---------------

For W_i to be an effective control variable in IV estimation, the following condition must hold

a) E(u_i) = 0

b) E(u_i|Z_i,W_i) = E(u_i|W_i)

c) E(u_i u_j) ≠ 0

d) there must be an intercept in the regression

---------------

As explained in class, b) is correct.

Here are additional explanations:

The condition E(u_i|Z_i,W_i) = E(u_i|W_i) ensures that the instrument Z_i is valid after accounting for the control variable W_i. Put differently, once we control for W_i, any remaining correlation between Z_i and u_i is eliminated and our IV estimation will provide an unbiased estimate of the causal effect of X_i on Y_i.

28.05.2024:

Dear students,

As mentioned yesterday, the lecture on the 24th of June will take place online.

Please find below the link to the Big Blue Button Webroom and the password.

https://webroom.hrz.tu-chemnitz.de/gl/lin-wfo-7zm-nkv

Password: 097860

The next tutorial will take place on the 18th of June (7:30-9:00 AM).

Update 12:12 PM: I have uploaded the additional slides on "probability vs. likelihood".

08.05.2024:

Dear students,

You can find all relevant information about our project "Racial Discrimination in the US Mortgage Market" in the folder "Project HMDA" (project description, data (excel) file, code that prepares the dataset). I have decided not to upload the code for running the actual analysis yet (you can find some hints in the shared Google Docs Document though). I will provide my "Example Code" which includes some key regressions and interpretations in the next week. 

As stated in class, feel free to hand in your short report and I will give you feedback if you wish. However, this is not mandatory.

30.04.2024:

Dear students,

I have uploaded the R code for the mortgage application denial example we discussed in class.

Remark: You have to convert the variable 'deny' to numeric as the lm() function does not accept the dependent variable to be a factor variable. See also the explanations in the code file. 

Update (11:08 AM): I have also uploaded the code for R-Exercise 2 (Chapter Panel Data) of the Problem Set. 

16.04.2024:

Dear students,

I have uploaded the updated slides for the lecture on panel data (it now includes the solution slides). 

Below, I have summarized the key points I mentioned in Monday's lecture regarding your questions, etc.

When is the effect of an independent variable constant and what is the role of time dummies in that context?

If we do not include interaction terms between the independent variables (such as education) and the time dummies, the model assumes that the effects of these independent variables on fertility are constant across all time periods (independent if we include time dummies or not). Including time fixed effects (year dummies) controls for the overall time trend in fertility. It allows us to capture any systematic differences in fertility across different years. Including interaction terms between education and the time dummies allows us to examine whether the effect of education on fertility varies across different time periods.

Fixed effects vs. random effects: More intuition (donut example, figure on slide 59)

Random effects exploit both "within" and "between" variation. The RE estimate is a weighted average of the within (fixed effects) and between estimates. I have uploaded additional graphical illustrations in the "Notes" folder. (The between estimator is obtained by estimating equation (2) on slide 29 (the "means over time") with OLS.)

14.04.2024:

Dear students,

I have uploaded the slides for tomorrow's lecture on panel data.

08.04.2024:

Dear students,

I have uploaded the updated Revision Quiz as well as the respective solution file in the folder "Problem Sets". The next lecture on the 15th of April will be about panel data analysis.

Please note that the tutorials will only take place after announcement. There is no tutorial tomorrow. To ensure that you have enough practice, I will include R practice sessions and general exercises in the Monday lectures.

12.03.2024:

Dear students,

I have created a "Google Docs" text document. During the R tutorial, I will occasionally post R code in this document that you can copy if a step is too fast for you. This is to ensure that nobody loses track. Of course, you can also address me directly during the exercise if you have any questions. Please note that occasionally some special characters such as "~" may not copy correctly, so you may need to manually enter them again in R. You are authorized to edit the document to add questions about your code, including the corresponding code section if necessary. However, please do not delete my contributions. The complete solution will usually be provided to you as an RScript on OPAL after the lecture/tutorial.

Here is the link to the Google Docs document: https://docs.google.com/document/d/1UnJBVpabqzfae-hHe0yEw4mKtFtuYoABq4nOBzxHxQA/edit?usp=sharing

--------------------------